Share your thoughts, 1 month free Claude Pro on usSee more

Math Reasoning on GSM8K (test)

95.9Accuracy

Agent-GWO

Updated 1mo ago

Evaluation Results

Method
Agent-GWO 2026.04	95.9	-	-	-	-	-
AoT 2026.04	95	-	-	-	-	-
ToT 2026.04	94.9	-	-	-	-	-
GPT-4-Turbo-0409 2024.05	94.5	-	-	-	-	-
AFlow 2026.04	93.5	-	-	-	-	-
MATH-SHEPHERD 2023.12	93.3	-	-	-	-	-
MATH-SHEPHERD 2023.12	93.2	-	-	-	-	-
GoT 2026.04	93.2	-	-	-	-	-
Agent-GWO 2026.04	92.8	-	-	-	-	-
ORM 2023.12	92.6	-	-	-	-	-
Self-Consistency + MATH-SHEPHERD 2023.12	92.5	-	-	-	-	-
Self-Consistency + MATH-SHEPHERD 2023.12	92.4	-	-	-	-	-
Self-Consistency + ORM 2023.12	92.4	-	-	-	-	-
Self-Consistency + ORM 2023.12	92	-	-	-	-	-
ORM 2023.12	91.8	-	-	-	-	-
AoT 2026.04	91.4	-	-	-	-	-
MATH-SHEPHERD 2023.12	90.9	-	-	-	-	-
aPSF 2026.04	90.87	-	-	-	-	-
GrIPS 2026.04	90.8	-	-	-	-	-
CriSPO 2026.04	90.13	-	-	-	-	-
CriSPO 2026.04	90.11	-	-	-	-	-
aPSF 2026.04	90.03	-	-	-	-	-
ORM 2023.12	90	-	-	-	-	-
ZERA 2026.04	89.98	-	-	-	-	-
CoT-SC/n=5 2026.04	89.9	-	-	-	-	-
Self-Consistency + MATH-SHEPHERD 2023.12	89.7	-	-	-	-	-
ProTeGi 2026.04	89.7	-	-	-	-	-
ZERA 2026.04	89.69	-	-	-	-	-
Self-Consistency + ORM 2023.12	89.6	-	-	-	-	-
ZERA 2026.04	89.55	-	-	-	-	-
CriSPO 2026.04	89.49	-	-	-	-	-
CriSPO 2026.04	89.4	-	-	-	-	-
GoT 2026.04	89.2	-	-	-	-	-
Agent-GWO 2026.04	89.1	-	-	-	-	-
AFlow 2026.04	89	-	-	-	-	-
Self-Refine 2026.04	88.9	-	-	-	-	-
APE 2026.04	88.83	-	-	-	-	-
OPRO 2026.04	88.7	-	-	-	-	-
AoT 2026.04	88.5	-	-	-	-	-
ProTeGi 2026.04	88.43	-	-	-	-	-
ToT 2026.04	88.4	-	-	-	-	-
Self-Consistency 2023.12	88.2	-	-	-	-	-
GrIPS 2026.04	88.16	-	-	-	-	-
Self-Consistency 2023.12	88	-	-	-	-	-
OPRO 2026.04	87.63	-	-	-	-	-
APE 2026.04	87.36	-	-	-	-	-
MAmmoTH2-8x7B-Plus 2024.05	86.4	-	-	-	-	-
Zero-shot CoT 2026.04	86.22	-	-	-	-	-
aPSF 2026.04	86.22	-	-	-	-	-
CoT-SC/n=5 2026.04	85.9	-	-	-	-	-
aPSF 2026.04	85.62	-	-	-	-	-
CoT 2026.04	85.4	-	-	-	-	-
OPRO 2026.04	85.25	-	-	-	-	-
CNTP + SC (Ours + SC) 2025.07	85.2	18,400	-	-	-	-
OPRO 2026.04	84.95	-	-	-	-	-
SC 2025.07	84.8	6,030	-	-	-	-
GoT 2026.04	84.8	-	-	-	-	-
MAmmoTH2-7B-Plus 2024.05	84.7	-	-	-	-	-
Self-Refine 2026.04	84.6	-	-	-	-	-
ProTeGi 2026.04	84.48	-	-	-	-	-
top-performing BoN 2026.05	84.47	-	-	-	-	-
GrIPS 2026.04	84.28	-	-	-	-	-
SLOP Hard selection 2026.05	84.27	-	-	-	-	-
Fixed weights (ω = 1) 2026.05	84.27	-	-	-	-	-
ZERA 2026.04	84.15	-	-	-	-	-
MAmmoTH2-8B-Plus 2024.05	84.1	-	-	-	-	-
AFlow 2026.04	84	-	-	-	-	-
SLOP sampling 2026.05	83.91	-	-	-	-	-
Diagonal Covariance hard selection 2026.05	83.77	-	-	-	-	-
Inverse Covariance hard selection 2026.05	83.67	-	-	-	-	-
ToT 2026.04	83.6	-	-	-	-	-
GrIPS 2026.04	83.55	-	-	-	-	-
CoT 2026.04	83.5	-	-	-	-	-
APE 2026.04	83.48	-	-	-	-	-
Dense w/o 2026.06	82.87	-	-	-	-	-
Zero-shot CoT 2026.04	82.81	-	-	-	-	-
Self-Consistency 2023.12	82.6	-	-	-	-	-
Beam Search 2025.07	82.3	740	-	-	-	-
CNTP (Ours*) 2025.07	81.6	460	-	-	-	-
Baseline reference sampling 2026.05	81.51	-	-	-	-	-
ProTeGi 2026.04	81.07	-	-	-	-	-
APE 2026.04	80.2	-	-	-	-	-
CoT-SC/n=5 2026.04	80	-	-	-	-	-
Greedy Decoding 2025.07	79.8	150	-	-	-	-
Stochastic Decoding* 2025.07	79.4	150	-	-	-	-
Self-Refine 2026.04	79	-	-	-	-	-
Minerva 2022.06	78.5	-	-	-	-	-
CoT 2026.04	77.5	-	-	-	-	-
InherNet 2026.02	77.19	-	-	-	-	-
SLOP Hard selection 2026.05	76.93	-	-	-	-	-
Student (Self-Distillation) 2026.02	76.88	-	-	-	-	-
Diagonal Covariance hard selection 2026.05	76.49	-	-	-	-	-
Teacher 2026.02	76.34	-	-	-	-	-
Inverse Covariance hard selection 2026.05	76.16	-	-	-	-	-
SLOP sampling 2026.05	76.05	-	-	-	-	-
MAmmoTH2-34B 2024.05	75.6	-	-	-	-	-
top-performing BoN 2026.05	75.51	-	-	-	-	-
MAmmoTH2-8x7B 2024.05	75.4	-	-	-	-	-
PaLM 540B maj1@40 (Published SOTA) 2022.06	74.4	-	-	-	-	-
CoT + Self-Consistency 2022.10	74.4	-	-	-	-	-

Showing 100 of 320 rows